With the large-scale development of information technologies, computers are widely applied in various fields. Faults in a computer may generally include a software fault, a hardware fault, an operation (configuration) fault, and other faults. A hardware fault has characteristics such as difficulty in reproduction, determining mainly depending on personnel experience, difficulty in locating a fault when an error occurs, the need for multiple times of insertion and removal/replacement, and the like. Therefore, a hardware fault, for example, a fault that occurs in memory, a processor, an input output (IO) device or the like, is usually the most difficult to process.
Generally, a hardware fault causes an uncorrectable error (Uncorrectable error) on a computer. An uncorrectable error not only can cause a service interruption in the computer and reduce an operation time of the computer, but also can even cause a breakdown event. In the prior art, a fault in a computer is mainly processed using the following method: When an uncorrectable error occurs in a system, a processor records error data and sends a notification to an operating system (OS). After receiving the notification, the OS captures the error data recorded by the processor and prints the error data, so that a user analyzes, locates a fault and makes a recovery from the fault.
In the prior art, an OS is needed to implement capturing of error data. However, once a severe uncorrectable error occurs in a computer and causes the computer to crash (in the present invention, a crash of a computer refers to that a black screen occurs in the computer, no input is accepted using an input device such as a mouse or keyboard of the computer, and a processor of the computer cannot execute a computer instruction), the OS can no longer work and cannot capture error data in the computer, making it difficult to analyze, process a fault and make a recovery from the fault.